moonlight-lupin/agent-skills / agent-ops/log-analyzer
log-analyzer
VerifiedUse when parsing agent log files to identify error patterns, rate limit hits, timeout clusters, tool failures, and component-level error counts. Produces a structured anomaly report. Cron-compatible — silent if no issues, alert digest if anomalies found.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 11.7 KB |
| references/anomaly-types.md | markdown | 4.3 KB |
| references/hermes-profile-audit.md | markdown | 3.9 KB |
| references/log-formats.md | markdown | 3.4 KB |
| references/state-failure-monitor.md | markdown | 4.5 KB |
| scripts/analyze_logs.pyexecutable | script | 23.1 KB |
| scripts/hermes_log_window.py | script | 4.2 KB |
| scripts/state_failures.pyexecutable | script | 8.8 KB |
| tests/test_analyze_logs.py | script | 8.1 KB |
| tests/test_state_failures.py | script | 9.0 KB |
references/hermes-profile-audit.md
View on GitHub# Auditing a Hermes Profile's Logs and Skill Usage
When you need a windowed performance/error/skill-usage report for a specific
Hermes profile (e.g. the Jing profile cron monitor), the agent.log completion
lines do NOT carry the skill-name argument — they only log
`tool skill_view completed (0.05s, 13890 chars)`. To get skill names and a
full tool-call breakdown, query the profile's own `state.db`.
## Where profile logs live
```
/root/.hermes/logs/ # default (MH) profile
/root/.hermes/profiles/<name>/logs/ # named profile
├── agent.log # INFO+ — tool completions, turn ends, API calls
├── errors.log # WARNING+ — deduped error stream
└── gateway.log # gateway lifecycle, telegram reconnects, MCP
```
Old/stale copies may exist at `/root/.hermes/profiles/<name>/.hermes-<name>/logs/`
— ignore those; the active logs are under `profiles/<name>/logs/`.
## The timestamp gotcha (why `--since` silently no-ops)
Hermes Python-logging lines look like:
```
2026-07-06 12:30:45,123 INFO hermes_cli.plugins: Plugin 'browser' ...
```
`analyze_logs.py` recognizes `YYYY-MM-DD HH:MM:SS LEVEL …` and ISO-8601, but
NOT the comma-millisecond + space-level form above. When the regex misses,
lines fall to the unstructured fallback (`timestamp=null`). `--since` cannot
filter null-timestamp lines, so the **entire unrotated file is scanned** and
every cluster reports `First: unknown / Last: unknown`.
Symptoms: error/traceback counts wildly exceed the true windowed count; the
error timeline says "No timestamped errors detected"; crashes are inflated
because each stack-trace continuation line counts as a separate crash.
**Fix:** run a small custom parser that handles the comma-ms format, applies
the window, and de-duplicates multiline tracebacks (count `Traceback (most
recent call last)` first-lines, not continuation frames). See
`scripts/hermes_log_window.py` for a ready-to-run windowed counter.
## Getting skill-usage and tool-call counts
The profile's `state.db` (SQLite) stores full tool-call arguments in the
`messages` table:
```python
import sqlite3, json
from collections import Counter
from datetime import datetime, timedelta
DB = f"/root/.hermes/profiles/{profile}/state.db"
END = datetime(2026, 7, 18, 15, 48, 27)
START = END - timedelta(days=4)
start_e, end_e = START.timestamp(), END.timestamp()
con = sqlite3.connect(f"file:{DB}?mode=ro", uri=True)
con.row_factory = sqlite3.Row
rows = con.execute("""
SELECT tool_name, tool_calls, timestamp
FROM messages
WHERE timestamp >= ? AND timestamp <= ?
""", (start_e, end_e)).fetchall()
tool_counts = Counter()
skill_view_targets = Counter()
for r in rows:
tc = r["tool_calls"]
if tc:
calls = json.loads(tc) if isinstance(tc, str) else tc
if isinstance(calls, dict): calls = [calls]
for c in calls:
name = c.get("name") or c.get("function", {}).get("name")
args = c.get("arguments") or c.get("function", {}).get("arguments") or {}
if isinstance(args, str):
try: args = json.loads(args)
except: pass
if name:
tool_counts[name] += 1
if name == "skill_view" and isinstance(args, dict):
sn = args.get("name")
if sn: skill_view_targets[sn] += 1
elif r["tool_name"]:
tool_counts[r["tool_name"]] += 1
```
`messages.timestamp` is a Unix epoch (REAL), not ISO — filter with
`timestamp >= start_epoch AND timestamp <= end_epoch`.
## Distinguishing profile-interactive sessions from cron sessions
Cron jobs that operate on a profile (e.g. a memory-curator cron targeting the
Jing profile) write their session IDs into `agent.log` but those sessions
live in the **default profile's** state.db, not the target profile's. If a
session ID from `agent.log` is missing from the profile's `state.db`, it was
a cross-profile cron, not the user's interactive use — exclude it from the
skill-usage count.Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |